Wavelet Mean Reversion Server

Overview

The Wavelet Mean Reversion Server fetches market data, analyzes returns or signals with wavelet decomposition, and generates synthetic data for experiments.

Connection

Add this server to your MCP client configuration.

Cursor
Bearer auth

{
  "mcpServers": {
    "wavelet-mean-reversion": {
      "url": "https://wavelet-mean-reversion-production.up.railway.app/mcp/waveletmeanreversion"
    }
  }
}

{
  "mcpServers": {
    "wavelet-mean-reversion": {
      "url": "https://wavelet-mean-reversion-production.up.railway.app/mcp/waveletmeanreversion",
      "headers": {
        "Authorization": "Bearer ${env:WAVELET_MEAN_REVERSION_MCP_TOKEN}"
      }
    }
  }
}

Restart the client after changing MCP configuration. The server tools appear automatically after the connection is established.

Transport

Property	Value
Protocol	MCP over Streamable HTTP
MCP URL	`https://wavelet-mean-reversion-production.up.railway.app/mcp/waveletmeanreversion`
Health URL	`https://wavelet-mean-reversion-production.up.railway.app/health`
MCP path	`/mcp/waveletmeanreversion`
Request envelope	`{"request": {...}}`
Auth	Optional bearer token when enabled for the endpoint

Best-Fit Workflows

Fetch Yahoo market data.
Compute returns for a ticker.
Run wavelet decomposition and mean reversion analysis.
Generate synthetic mean-reverting or trending data.
Backtest simple wavelet mean-reversion strategies.

Recommended Tools

yahoo-data-fetcher-fetch
yahoo-data-fetcher-fetch-returns
wavelet-analyzer-analyze
wavelet-analysis-result-summary
wavelet-analysis-result-get-statistics-dataframe

Tools

Tool	Description	Returns
`main`	Main example demonstrating wavelet mean reversion analysis and trading strategy.	`Any`
`yahoo-data-fetcher-fetch`	Fetches historical stock data from Yahoo Finance for one or more tickers. Params: ticker (str or list): stock symbol(s); start_date (str, optional): start date; end_date (str, optional): end date; period (str, default=‘max’): time period; interval (str, default=‘1d’): data frequency	`Any`
`wavelet-mean-reversion-strategy-generate-signals`	Generate trading signals.	`Any`
`wavelet-analyzer-decompose`	Decomposes a signal into wavelet components for mean reversion analysis. Params: signal: np.ndarray (input signal to decompose)	`Any`
`wavelet-analyzer-analyze`	Analyzes a signal using wavelet decomposition to detect mean reversion patterns. Outputs analysis results for further processing in trading strategies. Params: signal: np.ndarray - input signal data to analyze (no default provided)	`Any`
`yahoo-data-fetcher-fetch-returns`	Fetches historical stock returns from Yahoo Finance for a given ticker symbol, supporting optional date ranges, return period, and log return calculation. Params: ticker (str), start_date (str, optional), end_date (str, optional), period (str, default=‘max’), log_returns (bool, default=True)	`Any`
`yahoo-data-fetcher-fetch-multiple-series`	Fetches historical stock price data for multiple tickers from Yahoo Finance, returning the specified column over the given date range. Params: tickers (list of str), start_date (str, optional), end_date (str, optional), period (str, default=‘max’), column (str, default=‘Close’)	`Any`
`wavelet-mean-reversion-strategy-backtest-simple`	Simple backtest of the strategy.	`Any`
`config-from-dict`	Create configuration from dictionary.	`Any`
`wavelet-mean-reversion-strategy-fit`	Fit the strategy by analyzing the signal.	`Any`
`synthetic-data-generator-generate-mean-reverting`	Generates synthetic mean-reverting time series data using Ornstein-Uhlenbeck process parameters. Params: N (default: 1000): number of data points; theta (default: 0.1): mean reversion speed; mu (default: 0.0): long-term mean; sigma (default: 1.0): volatility; seed (default: 42): random seed for reproducibility	`Any`
`synthetic-data-generator-generate-trending`	Generates synthetic trending financial data using a mean-reverting model with configurable parameters for length, drift, and volatility. Params: Inputs: N (int, default 1000), drift (float, default 0.01), volatility (float, default 1.0), seed (Optional[int], default 42)	`Any`
`wavelet-visualizer-plot-trading-signals`	Plot trading signals on price chart.	`Any`
`wavelet-visualizer-plot-backtest-results`	Plot backtest performance.	`Any`
`synthetic-data-generator-generate-complex-signal`	Generates a synthetic complex signal for wavelet analysis, producing N data points with optional random seed and component breakdown. Params: N (int, default=2000): number of data points; seed (int, optional, default=42): random seed; include_components (bool, default=False): whether to return individual signal components	`Any`
`synthetic-data-generator-add-noise`	Adds random noise to a signal array using a configurable noise level and optional seed for reproducibility. Params: signal (np.ndarray): input signal to add noise to; noise_level (float, default 0.1): amplitude of added noise; seed (Optional[int], default None): random seed for reproducibility	`Any`
`config-to-dict`	Convert configuration to dictionary.	`Any`
`wavelet-visualizer-plot-complete-analysis`	Create comprehensive analysis visualization.	`Any`
`wavelet-visualizer-plot-acf-analysis`	Plot autocorrelation function for each scale.	`Any`
`wavelet-visualizer-plot-monte-carlo-results`	Plot Monte Carlo simulation results.	`Any`
`wavelet-decomposition-result-get-scale-signal`	Extracts the signal for a specific wavelet scale from decomposition results. Params: scale_name (str): name of the wavelet scale to extract signal from (e.g., ‘D1’, ‘D2’)	`Any`
`trading-signal-to-dataframe`	Convert signals to DataFrame.	`Any`
`wavelet-analysis-result-summary`	Generates a concise summary of wavelet analysis results, providing key metrics and insights from the wavelet decomposition and mean reversion analysis. Params: none	`Any`
`wavelet-analysis-result-get-statistics-dataframe`	Retrieves statistical data from wavelet analysis results as a pandas DataFrame, providing summary metrics for mean reversion analysis. Params: none	`Any`
`main-main`	Run all examples.	`Any`
`validate-time-series`	Validates and converts time series data to a 1D numpy array, ensuring it has at least 4 points and contains no NaN or inf values. Params: data (Union[np.ndarray, pd.Series]): Time series data to validate	`np.ndarray`
`compare-with-bollinger-bands`	Compares a wavelet-based mean-reverting component with traditional Bollinger Bands, returning metrics for both approaches. Params: signal: Original signal array; wavelet_mr_component: Wavelet-derived mean-reverting component; window: Bollinger Band window (default 20); num_std: Std deviation multiplier (default 2.0)	`Dict[str, np.ndarray]`
`example-synthetic-data`	Example using synthetic data.	`Any`
`example-real-market-data`	Example using real market data.	`Any`
`example-monte-carlo-backtest`	Example with Monte Carlo backtesting.	`Any`
`example-multiple-tickers`	Example analyzing multiple tickers.	`Any`
`example-custom-configuration`	Example using custom configuration.	`Any`
`default-config`	Preset factory: create Config via ObjectStore (handles) with default deps (none); returns a handle.	`object`
`monte-carlo-default`	Preset factory: create MonteCarloBacktester via ObjectStore (handles) with default deps (none); returns a handle.	`object`
`statistical-analyzer-default`	Preset factory: create StatisticalAnalyzer via ObjectStore (handles) with default deps (none); returns a handle.	`object`
`wavelet-analyzer-default`	Preset factory: create WaveletAnalyzer via ObjectStore (handles) with default deps (none); returns a handle.	`object`
`wavelet-strategy-default`	Preset factory: create WaveletMeanReversionStrategy via ObjectStore (handles) with default deps (analyzer); returns a handle.	`object`
`wavelet-visualizer-default`	Preset factory: create WaveletVisualizer via ObjectStore (handles) with default deps (none); returns a handle.	`object`
`yahoo-fetcher-default`	Preset factory: create YahooDataFetcher via ObjectStore (handles) with default deps (none); returns a handle.	`object`

Examples

Fetch returns for one ticker

{
  "request": {
    "ticker": "AAPL",
    "start_date": "2024-01-01",
    "end_date": "2024-06-30",
    "period": "1y",
    "log_returns": true
  }
}

Analyze a signal

{
  "request": {
    "signal": [
      101.0,
      100.5,
      100.9,
      100.1,
      99.8,
      100.2
    ]
  }
}

Notes

Instance-style tools can use handles when you need configured analyzers or strategies.
Use Parallax ExtremeHurst when you need the ExtremeHurst signal engine rather than wavelet decomposition.

Client setup

Configure this endpoint in Cursor, Claude Desktop, or a generic MCP client.

Shared tools

Use health, result, artifact, environment, and table helper tools.

Other Servers

Dataloader

Market-data retrieval, TA-Lib indicators, and dataframe exports.

Fama-French Replicate

Official and replicated Fama-French factors plus loadings and alpha estimation.

Statistical Factor Models

Stock-Watson, complete-panel, and dynamic statistical factor extraction.

Jump Models

JumpModel and SparseJumpModel regime fitting, online prediction, and backtesting.

Parallax ExtremeHurst

ExtremeHurst signal generation from OHLCV data.

EP Ratio Screener

Fundamental stock screening based on earnings yield and balance-sheet quality.

Volatility Scaling Lab

Volatility targeting, EWMA volatility, Monte Carlo bands, and risk diagnostics.

Overview

MCP Setup

Dataloader Server

Fama-French Replicate Server

Statistical Factor Models Server

Jump Models Server